2 ''' Copyright Neil Brown and others.
3 ''' This program is free software; you can redistribute it and/or modify
4 ''' it under the terms of the GNU General Public License as published by
5 ''' the Free Software Foundation; either version 2 of the License, or
6 ''' (at your option) any later version.
7 ''' See file COPYING in distribution for details.
10 mdadm \- manage MD devices
16 .BI mdadm " [mode] <raiddevice> [options] <component-devices>"
19 RAID devices are virtual devices created from two or more
20 real block devices. This allows multiple devices (typically disk
21 drives or partitions thereof) to be combined into a single device to
22 hold (for example) a single filesystem.
23 Some RAID levels include redundancy and so can survive some degree of
26 Linux Software RAID devices are implemented through the md (Multiple
27 Devices) device driver.
29 Currently, Linux supports
45 is not a Software RAID mechanism, but does involve
48 each device is a path to one common physical storage device.
51 is also not true RAID, and it only involves one device. It
52 provides a layer over a true device that can be used to inject faults.
55 '''is a program that can be used to create, manage, and monitor
57 '''such it provides a similar set of functionality to the
60 '''The key differences between
67 '''is a single program and not a collection of programs.
70 '''can perform (almost) all of its functions without having a
71 '''configuration file and does not use one by default. Also
73 '''helps with management of the configuration
77 '''can provide information about your arrays (through Query, Detail, and Examine)
87 '''configuration file, at all. It has a different configuration file
88 '''with a different format and a different purpose.
91 mdadm has several major modes of operation:
94 Assemble the parts of a previously created
95 array into an active array. Components can be explicitly given
96 or can be searched for.
98 checks that the components
99 do form a bona fide array, and can, on request, fiddle superblock
100 information so as to assemble a faulty array.
104 Build an array that doesn't have per-device superblocks. For these
107 cannot differentiate between initial creation and subsequent assembly
108 of an array. It also cannot perform any checks that appropriate
109 devices have been requested. Because of this, the
111 mode should only be used together with a complete understanding of
116 Create a new array with per-device superblocks.
118 '''in several step create-add-add-run or it can all happen with one command.
121 .B "Follow or Monitor"
122 Monitor one or more md devices and act on any state changes. This is
123 only meaningful for raid1, 4, 5, 6, 10 or multipath arrays as
124 only these have interesting state. raid0 or linear never have
125 missing, spare, or failed drives, so there is nothing to monitor.
129 Grow (or shrink) an array, or otherwise reshape it in some way.
130 Currently supported growth options including changing the active size
131 of component devices in RAID level 1/4/5/6 and changing the number of
132 active devices in RAID1/5/6.
135 .B "Incremental Assembly"
136 Add a single device to an appropriate array. If the addition of the
137 device makes the array runnable, the array will be started.
138 This provides a convenient interface to a
140 system. As each device is detected,
142 has a chance to include it in some array as appropriate.
146 This is for doing things to specific components of an array such as
147 adding new spares and removing faulty devices.
151 This is an 'everything else' mode that supports operations on active
152 arrays, operations on component devices such as erasing old superblocks, and
153 information gathering operations.
154 '''This mode allows operations on independent devices such as examine MD
155 '''superblocks, erasing old superblocks and stopping active arrays.
159 This mode does not act on a specific device or array, but rather it
160 requests the Linux Kernel to activate any auto-detected arrays.
163 .SH Options for selecting a mode are:
166 .BR \-A ", " \-\-assemble
167 Assemble a pre-existing array.
170 .BR \-B ", " \-\-build
171 Build a legacy array without superblocks.
174 .BR \-C ", " \-\-create
178 .BR \-F ", " \-\-follow ", " \-\-monitor
184 .BR \-G ", " \-\-grow
185 Change the size or shape of an active array.
188 .BR \-I ", " \-\-incremental
189 Add a single device into an appropriate array, and possibly start the array.
193 Request that the kernel starts any auto-detected arrays. This can only
196 is compiled into the kernel \(em not if it is a module.
197 Arrays can be auto-detected by the kernel if all the components are in
198 primary MS-DOS partitions with partition type
200 In-kernel autodetect is not recommended for new installations. Using
202 to detect and assemble arrays \(em possibly in an
204 \(em is substantially more flexible and should be preferred.
207 If a device is given before any options, or if the first option is
212 then the MANAGE mode is assume.
213 Anything other than these will cause the
217 .SH Options that are not mode-specific are:
220 .BR \-h ", " \-\-help
221 Display general help message or, after one of the above options, a
222 mode specific help message.
226 Display more detailed help about command line parsing and some commonly
230 .BR \-V ", " \-\-version
231 Print version information for mdadm.
234 .BR \-v ", " \-\-verbose
235 Be more verbose about what is happening. This can be used twice to be
237 The extra verbosity currently only affects
238 .B \-\-detail \-\-scan
240 .BR "\-\-examine \-\-scan" .
243 .BR \-q ", " \-\-quiet
244 Avoid printing purely informative messages. With this,
246 will be silent unless there is something really important to report.
249 .BR \-b ", " \-\-brief
250 Be less verbose. This is used with
258 gives an intermediate level of verbosity.
261 .BR \-f ", " \-\-force
262 Be more forceful about certain operations. See the various modes of
263 the exact meaning of this option in different contexts.
266 .BR \-c ", " \-\-config=
267 Specify the config file. Default is to use
268 .BR /etc/mdadm.conf ,
269 or if that is missing, then
270 .BR /etc/mdadm/mdadm.conf .
271 If the config file given is
273 then nothing will be read, but
275 will act as though the config file contained exactly
276 .B "DEVICE partitions"
279 to find a list of devices to scan.
282 is given for the config file, then
284 will act as though the config file were empty.
287 .BR \-s ", " \-\-scan
290 for missing information.
291 In general, this option gives
293 permission to get any missing information, like component devices,
294 array devices, array identities, and alert destination from the
296 .BR /etc/mdadm.conf .
297 One exception is MISC mode when using
303 says to get a list of array devices from
307 .B \-e ", " \-\-metadata=
308 Declare the style of superblock (raid metadata) to be used. The
311 and to guess for other operations.
312 The default can be overridden by setting the
321 .IP "0, 0.90, default"
322 Use the original 0.90 format superblock. This format limits arrays to
323 28 componenet devices and limits component devices of levels 1 and
324 greater to 2 terabytes.
325 .IP "1, 1.0, 1.1, 1.2"
326 Use the new version-1 format superblock. This has few restrictions.
327 The different subversion store the superblock at different locations
328 on the device, either at the end (for 1.0), at the start (for 1.1) or
329 4K from the start (for 1.2).
334 This will override any
336 setting in the config file and provides the identify of the host which
337 should be considered the home for any arrays.
339 When creating an array, the
341 will be recorded in the superblock. For version-1 superblocks, it will
342 be prefixed to the array name. For version-0.90 superblocks part of
343 the SHA1 hash of the hostname will be stored in the later half of the
346 When reporting information about an array, any array which is tagged
347 for the given homehost will be reported as such.
349 When using Auto-Assemble, only arrays tagged for the given homehost
352 .SH For create, build, or grow:
355 .BR \-n ", " \-\-raid\-devices=
356 Specify the number of active devices in the array. This, plus the
357 number of spare devices (see below) must equal the number of
359 (including "\fBmissing\fP" devices)
360 that are listed on the command line for
362 Setting a value of 1 is probably
363 a mistake and so requires that
365 be specified first. A value of 1 will then be allowed for linear,
366 multipath, raid0 and raid1. It is never allowed for raid4 or raid5.
368 This number can only be changed using
370 for RAID1, RAID5 and RAID6 arrays, and only on kernels which provide
374 .BR \-x ", " \-\-spare\-devices=
375 Specify the number of spare (eXtra) devices in the initial array.
376 Spares can also be added
377 and removed later. The number of component devices listed
378 on the command line must equal the number of raid devices plus the
379 number of spare devices.
383 .BR \-z ", " \-\-size=
384 Amount (in Kibibytes) of space to use from each drive in RAID1/4/5/6.
385 This must be a multiple of the chunk size, and must leave about 128Kb
386 of space at the end of the drive for the RAID superblock.
387 If this is not specified
388 (as it normally is not) the smallest drive (or partition) sets the
389 size, though if there is a variance among the drives of greater than 1%, a warning is
392 This value can be set with
394 for RAID level 1/4/5/6. If the array was created with a size smaller
395 than the currently active drives, the extra space can be accessed
398 The size can be given as
400 which means to choose the largest size that fits on all current drives.
403 .BR \-c ", " \-\-chunk=
404 Specify chunk size of kibibytes. The default is 64.
408 Specify rounding factor for linear array (==chunk size)
411 .BR \-l ", " \-\-level=
412 Set raid level. When used with
414 options are: linear, raid0, 0, stripe, raid1, 1, mirror, raid4, 4,
415 raid5, 5, raid6, 6, raid10, 10, multipath, mp, faulty. Obviously some of these are synonymous.
419 only linear, stripe, raid0, 0, raid1, multipath, mp, and faulty are valid.
421 Not yet supported with
425 .BR \-p ", " \-\-layout=
426 This option configures the fine details of data layout for raid5,
427 and raid10 arrays, and controls the failure modes for
430 The layout of the raid5 parity block can be one of
431 .BR left\-asymmetric ,
432 .BR left\-symmetric ,
433 .BR right\-asymmetric ,
434 .BR right\-symmetric ,
435 .BR la ", " ra ", " ls ", " rs .
437 .BR left\-symmetric .
439 When setting the failure mode for
442 .BR write\-transient ", " wt ,
443 .BR read\-transient ", " rt ,
444 .BR write\-persistent ", " wp ,
445 .BR read\-persistent ", " rp ,
447 .BR read\-fixable ", " rf ,
448 .BR clear ", " flush ", " none .
450 Each mode can be followed by a number which is used as a period
451 between fault generation. Without a number, the fault is generated
452 once on the first relevant request. With a number, the fault will be
453 generated after that many request, and will continue to be generated
454 every time the period elapses.
456 Multiple failure modes can be current simultaneously by using the
458 option to set subsequent failure modes.
460 "clear" or "none" will remove any pending or periodic failure modes,
461 and "flush" will clear any persistent faults.
463 To set the parity with
465 the level of the array ("faulty")
466 must be specified before the fault mode is specified.
468 Finally, the layout options for RAID10 are one of 'n', 'o' or 'f' followed
469 by a small number. The default is 'n2'.
472 signals 'near' copies. Multiple copies of one data block are at
473 similar offsets in different devices.
476 signals 'offset' copies. Rather than the chunks being duplicated
477 within a stripe, whole stripes are duplicated but are rotated by one
478 device so duplicate blocks are on different devices. Thus subsequent
479 copies of a block are in the next drive, and are one chunk further
484 (multiple copies have very different offsets). See md(4) for more
485 detail about 'near' and 'far'.
487 The number is the number of copies of each datablock. 2 is normal, 3
488 can be useful. This number can be at most equal to the number of
489 devices in the array. It does not need to divide evenly into that
490 number (e.g. it is perfectly legal to have an 'n2' layout for an array
491 with an odd number of devices).
497 (thus explaining the p of
501 .BR \-b ", " \-\-bitmap=
502 Specify a file to store a write-intent bitmap in. The file should not
505 is also given. The same file should be provided
506 when assembling the array. If the word
508 is given, then the bitmap is stored with the metadata on the array,
509 and so is replicated on all devices. If the word
513 mode, then any bitmap that is present is removed.
515 To help catch typing errors, the filename must contain at least one
516 slash ('/') if it is a real file (not 'internal' or 'none').
518 Note: external bitmaps are only known to work on ext2 and ext3.
519 Storing bitmap files on other filesystems may result in serious problems.
522 .BR \-\-bitmap\-chunk=
523 Set the chunksize of the bitmap. Each bit corresponds to that many
524 Kilobytes of storage.
525 When using a file based bitmap, the default is to use the smallest
526 size that is atleast 4 and requires no more than 2^21 chunks.
529 bitmap, the chunksize is automatically determined to make best use of
534 .BR \-W ", " \-\-write\-mostly
535 subsequent devices lists in a
540 command will be flagged as 'write-mostly'. This is valid for RAID1
541 only and means that the 'md' driver will avoid reading from these
542 devices if at all possible. This can be useful if mirroring over a
546 .BR \-\-write\-behind=
547 Specify that write-behind mode should be enabled (valid for RAID1
548 only). If an argument is specified, it will set the maximum number
549 of outstanding writes allowed. The default value is 256.
550 A write-intent bitmap is required in order to use write-behind
551 mode, and write-behind is only attempted on drives marked as
555 .BR \-\-assume\-clean
558 that the array pre-existed and is known to be clean. It can be useful
559 when trying to recover from a major failure as you can be sure that no
560 data will be affected unless you actually write to the array. It can
561 also be used when creating a RAID1 or RAID10 if you want to avoid the
562 initial resync, however this practice \(em while normally safe \(em is not
563 recommended. Use this ony if you really know what you are doing.
566 .BR \-\-backup\-file=
569 is used to increase the number of
570 raid-devices in a RAID5 if there are no spare devices available.
571 See the section below on RAID_DEVICE CHANGES. The file should be
572 stored on a separate device, not on the raid array being reshaped.
575 .BR \-N ", " \-\-name=
578 for the array. This is currently only effective when creating an
579 array with a version-1 superblock. The name is a simple textual
580 string that can be used to identify array components when assembling.
586 run the array, even if some of the components
587 appear to be active in another array or filesystem. Normally
589 will ask for confirmation before including such components in an
590 array. This option causes that question to be suppressed.
593 .BR \-f ", " \-\-force
596 accept the geometry and layout specified without question. Normally
598 will not allow creation of an array with only one device, and will try
599 to create a raid5 array with one missing drive (as this makes the
600 initial resync work faster). With
603 will not try to be so clever.
606 .BR \-a ", " "\-\-auto{=no,yes,md,mdp,part,p}{NN}"
607 Instruct mdadm to create the device file if needed, possibly allocating
608 an unused minor number. "md" causes a non-partitionable array
609 to be used. "mdp", "part" or "p" causes a partitionable array (2.6 and
610 later) to be used. "yes" requires the named md device to have
611 a 'standard' format, and the type and minor number will be determined
612 from this. See DEVICE NAMES below.
614 The argument can also come immediately after
619 is not given on the command line or in the config file, then
625 is also given, then any
627 entries in the config file will override the
629 instruction given on the command line.
631 For partitionable arrays,
633 will create the device file for the whole array and for the first 4
634 partitions. A different number of partitions can be specified at the
635 end of this option (e.g.
637 If the device name ends with a digit, the partition names add a 'p',
638 and a number, e.g. "/dev/home1p3". If there is no
639 trailing digit, then the partition names just have a number added,
640 e.g. "/dev/scratch3".
642 If the md device name is in a 'standard' format as described in DEVICE
643 NAMES, then it will be created, if necessary, with the appropriate
644 number based on that name. If the device name is not in one of these
645 formats, then a unused minor number will be allocated. The minor
646 number will be considered unused if there is no active array for that
647 number, and there is no entry in /dev for that number and with a
658 it will also create symlinks from
660 with names starting with
668 to enforce this even if it is suppressing
675 .BR \-u ", " \-\-uuid=
676 uuid of array to assemble. Devices which don't have this uuid are
680 .BR \-m ", " \-\-super\-minor=
681 Minor number of device that array was created for. Devices which
682 don't have this minor number are excluded. If you create an array as
683 /dev/md1, then all superblocks will contain the minor number 1, even if
684 the array is later assembled as /dev/md2.
686 Giving the literal word "dev" for
690 to use the minor number of the md device that is being assembled.
693 .M \-\-super\-minor=dev
694 will look for super blocks with a minor number of 0.
697 .BR \-N ", " \-\-name=
698 Specify the name of the array to assemble. This must be the name
699 that was specified when creating the array. It must either match
700 then name stored in the superblock exactly, or it must match
703 is added to the start of the given name.
706 .BR \-f ", " \-\-force
707 Assemble the array even if some superblocks appear out-of-date
711 Attempt to start the array even if fewer drives were given than were
712 present last time the array was active. Normally if not all the
713 expected drives are found and
715 is not used, then the array will be assembled but not started.
718 an attempt will be made to start it anyway.
722 This is the reverse of
724 in that it inhibits the started if array unless all expected drives
725 are present. This is only needed with
727 and can be used if you physical connections to devices are
728 not as reliable as you would like.
731 .BR \-a ", " "\-\-auto{=no,yes,md,mdp,part}"
732 See this option under Create and Build options.
735 .BR \-b ", " \-\-bitmap=
736 Specify the bitmap file that was given when the array was created. If
739 bitmap, there is no need to specify this when assembling the array.
742 .BR \-\-backup\-file=
745 was used to grow the number of raid-devices in a RAID5, and the system
746 crashed during the critical section, then the same
750 to allow possibly corrupted data to be restored.
753 .BR \-U ", " \-\-update=
754 Update the superblock on each device while assembling the array. The
755 argument given to this flag can be one of
769 option will adjust the superblock of an array what was created on a Sparc
770 machine running a patched 2.2 Linux kernel. This kernel got the
771 alignment of part of the superblock wrong. You can use the
772 .B "\-\-examine \-\-sparc2.2"
775 to see what effect this would have.
779 option will update the
781 field on each superblock to match the minor number of the array being
783 This can be useful if
785 reports a different "Preferred Minor" to
787 In some cases this update will be performed automatically
788 by the kernel driver. In particular the update happens automatically
789 at the first write to an array with redundancy (RAID level 1 or
790 greater) on a 2.6 (or later) kernel.
794 option will change the uuid of the array. If a UUID is given with the
796 option that UUID will be used as a new UUID and will
798 be used to help identify the devices in the array.
801 is given, a random UUID is chosen.
805 option will change the
807 of the array as stored in the superblock. This is only supported for
808 version-1 superblocks.
812 option will change the
814 as recorded in the superblock. For version-0 superblocks, this is the
815 same as updating the UUID.
816 For version-1 superblocks, this involves updating the name.
820 option will cause the array to be marked
822 meaning that any redundancy in the array (e.g. parity for raid5,
823 copies for raid1) may be incorrect. This will cause the raid system
824 to perform a "resync" pass to make sure that all redundant information
829 option allows arrays to be moved between machines with different
831 When assembling such an array for the first time after a move, giving
832 .B "\-\-update=byteorder"
835 to expect superblocks to have their byteorder reversed, and will
836 correct that order before assembling the array. This is only valid
837 with original (Version 0.90) superblocks.
841 option will correct the summaries in the superblock. That is the
842 counts of total, working, active, failed, and spare devices.
846 will rarely be of use. It applies to version 1.1 and 1.2 metadata
847 only (where the metadata is at the start of the device) and is only
848 useful when the component device has changed size (typically become
849 larger). The version 1 metadata records the amount of the device that
850 can be used to store data, so if a device in a version 1.1 or 1.2
851 array becomes larger, the metadata will still be visible, but the
852 extra space will not. In this case it might be useful to assemble the
854 .BR \-\-update=devicesize .
857 to determine the maximum usable amount of space on each device and
858 update the relevant field in the metadata.
861 .B \-\-auto\-update\-homehost
862 This flag is only meaning with auto-assembly (see discussion below).
863 In that situation, if no suitable arrays are found for this homehost,
865 will recan for any arrays at all and will assemble them and update the
866 homehost to match the current host.
872 hot-add listed devices.
876 re-add a device that was recently removed from an array.
879 .BR \-r ", " \-\-remove
880 remove listed devices. They must not be active. i.e. they should
881 be failed or spare devices. As well as the name of a device file
890 The first causes all failed device to be removed. The second causes
891 any device which is no longer connected to the system (i.e and open
894 to be removed. This will only succeed for devices that are spares or
895 have already been marked as failed.
898 .BR \-f ", " \-\-fail
899 mark listed devices as faulty.
900 As well as the name of a device file, the word
902 can be given. This will cause any device that has been detached from
903 the system to be marked as failed. It can then be removed.
911 Each of these options require that the first device list is the array
912 to be acted upon and the remainder are component devices to be added,
913 removed, or marked as fault. Several different operations can be
914 specified for different devices, e.g.
916 mdadm /dev/md0 \-\-add /dev/sda1 \-\-fail /dev/sdb1 \-\-remove /dev/sdb1
918 Each operation applies to all devices listed until the next
921 If an array is using a write-intent bitmap, then devices which have
922 been removed can be re-added in a way that avoids a full
923 reconstruction but instead just updated the blocks that have changed
924 since the device was removed. For arrays with persistent metadata
925 (superblocks) this is done automatically. For arrays created with
927 mdadm needs to be told that this device we removed recently with
930 Devices can only be removed from an array if they are not in active
931 use. i.e. that must be spares or failed devices. To remove an active
932 device, it must be marked as
939 .BR \-Q ", " \-\-query
940 Examine a device to see
941 (1) if it is an md device and (2) if it is a component of an md
943 Information about what is discovered is presented.
946 .BR \-D ", " \-\-detail
947 Print detail of one or more md devices.
950 .BR \-Y ", " \-\-export
953 output will be formatted as
955 pairs for easy import into the environment.
958 .BR \-E ", " \-\-examine
959 Print content of md superblock on device(s).
962 If an array was created on a 2.2 Linux kernel patched with RAID
963 support, the superblock will have been created incorrectly, or at
964 least incompatibly with 2.4 and later kernels. Using the
968 will fix the superblock before displaying it. If this appears to do
969 the right thing, then the array can be successfully assembled using
970 .BR "\-\-assemble \-\-update=sparc2.2" .
973 .BR \-X ", " \-\-examine\-bitmap
974 Report information about a bitmap file.
975 The argument is either an external bitmap file or an array component
976 in case of an internal bitmap.
980 start a partially built array.
983 .BR \-S ", " \-\-stop
984 deactivate array, releasing all resources.
987 .BR \-o ", " \-\-readonly
988 mark array as readonly.
991 .BR \-w ", " \-\-readwrite
992 mark array as readwrite.
995 .B \-\-zero\-superblock
996 If the device contains a valid md superblock, the block is
997 overwritten with zeros. With
999 the block where the superblock would be is overwritten even if it
1000 doesn't appear to be valid.
1003 .BR \-t ", " \-\-test
1008 is set to reflect the status of the device.
1011 .BR \-W ", " \-\-wait
1012 For each md device given, wait for any resync, recovery, or reshape
1013 activity to finish before returning.
1015 will return with success if it actually waited for every device
1016 listed, otherwise it will return failure.
1018 .SH For Incremental Assembly mode:
1020 .BR \-\-rebuild\-map ", " \-r
1021 Rebuild the map file
1022 .RB ( /var/run/mdadm/map )
1025 uses to help track which arrays are currently being assembled.
1028 .BR \-\-run ", " \-R
1029 Run any array assembled as soon as a minimal number of devices are
1030 available, rather than waiting until all expected devices are present.
1033 .BR \-\-scan ", " \-s
1034 Only meaningful with
1038 file for arrays that are being incrementally assembled and will try to
1039 start any that are not already started. If any such array is listed
1042 as requiring an external bitmap, that bitmap will be attached first.
1044 .SH For Monitor mode:
1046 .BR \-m ", " \-\-mail
1047 Give a mail address to send alerts to.
1050 .BR \-p ", " \-\-program ", " \-\-alert
1051 Give a program to be run whenever an event is detected.
1054 .BR \-y ", " \-\-syslog
1055 Cause all events to be reported through 'syslog'. The messages have
1056 facility of 'daemon' and varying priorities.
1059 .BR \-d ", " \-\-delay
1060 Give a delay in seconds.
1062 polls the md arrays and then waits this many seconds before polling
1063 again. The default is 60 seconds.
1066 .BR \-f ", " \-\-daemonise
1069 to run as a background daemon if it decides to monitor anything. This
1070 causes it to fork and run in the child, and to disconnect form the
1071 terminal. The process id of the child is written to stdout.
1074 which will only continue monitoring if a mail address or alert program
1075 is found in the config file.
1078 .BR \-i ", " \-\-pid\-file
1081 is running in daemon mode, write the pid of the daemon process to
1082 the specified file, instead of printing it on standard output.
1085 .BR \-1 ", " \-\-oneshot
1086 Check arrays only once. This will generate
1088 events and more significantly
1094 .B " mdadm \-\-monitor \-\-scan \-1"
1096 from a cron script will ensure regular notification of any degraded arrays.
1099 .BR \-t ", " \-\-test
1102 alert for every array found at startup. This alert gets mailed and
1103 passed to the alert program. This can be used for testing that alert
1104 message do get through successfully.
1110 .B mdadm \-\-assemble
1111 .I md-device options-and-component-devices...
1114 .B mdadm \-\-assemble \-\-scan
1115 .I md-devices-and-options...
1118 .B mdadm \-\-assemble \-\-scan
1122 This usage assembles one or more raid arrays from pre-existing components.
1123 For each array, mdadm needs to know the md device, the identity of the
1124 array, and a number of component-devices. These can be found in a number of ways.
1126 In the first usage example (without the
1128 the first device given is the md device.
1129 In the second usage example, all devices listed are treated as md
1130 devices and assembly is attempted.
1131 In the third (where no devices are listed) all md devices that are
1132 listed in the configuration file are assembled.
1134 If precisely one device is listed, but
1140 was given and identify information is extracted from the configuration file.
1142 The identity can be given with the
1146 option, can be found in the config file, or will be taken from the
1147 super block on the first component-device listed on the command line.
1149 Devices can be given on the
1151 command line or in the config file. Only devices which have an md
1152 superblock which contains the right identity will be considered for
1155 The config file is only used if explicitly named with
1157 or requested with (a possibly implicit)
1165 is not given, then the config file will only be used to find the
1166 identity of md arrays.
1168 Normally the array will be started after it is assembled. However if
1170 is not given and insufficient drives were listed to start a complete
1171 (non-degraded) array, then the array is not started (to guard against
1172 usage errors). To insist that the array be started in this case (as
1173 may work for RAID1, 4, 5, 6, or 10), give the
1177 If the md device does not exist, then it will be created providing the
1178 intent is clear. i.e. the name must be in a standard form, or the
1180 option must be given to clarify how and whether the device should be
1183 This can be useful for handling partitioned devices (which don't have
1184 a stable device number \(em it can change after a reboot) and when using
1185 "udev" to manage your
1187 tree (udev cannot handle md devices because of the unusual device
1188 initialisation conventions).
1190 If the option to "auto" is "mdp" or "part" or (on the command line
1191 only) "p", then mdadm will create a partitionable array, using the
1192 first free one that is not in use, and does not already have an entry
1193 in /dev (apart from numeric /dev/md* entries).
1195 If the option to "auto" is "yes" or "md" or (on the command line)
1196 nothing, then mdadm will create a traditional, non-partitionable md
1199 It is expected that the "auto" functionality will be used to create
1200 device entries with meaningful names such as "/dev/md/home" or
1201 "/dev/md/root", rather than names based on the numerical array number.
1203 When using this option to create a partitionable array, the device
1204 files for the first 4 partitions are also created. If a different
1205 number is required it can be simply appended to the auto option.
1206 e.g. "auto=part8". Partition names are created by appending a digit
1207 string to the device name, with an intervening "p" if the device name
1212 option is also available in Build and Create modes. As those modes do
1213 not use a config file, the "auto=" config option does not apply to
1221 and no devices are listed,
1223 will first attempt to assemble all the arrays listed in the config
1228 has been specified (either in the config file or on the command line),
1230 will look further for possible arrays and will try to assemble
1231 anything that it finds which is tagged as belonging to the given
1232 homehost. This is the only situation where
1234 will assemble arrays without being given specific device name or
1235 identify information for the array.
1239 finds a consistent set of devices that look like they should comprise
1240 an array, and if the superblock is tagged as belonging to the given
1241 home host, it will automatically choose a device name and try to
1242 assemble the array. If the array uses version-0.90 metadata, then the
1244 number as recorded in the superblock is used to create a name in
1248 If the array uses version-1 metadata, then the
1250 from the superblock is used to similarly create a name in
1252 The name will have any 'host' prefix stripped first.
1256 cannot find any array for the given host at all, and if
1257 .B \-\-auto\-update\-homehost
1260 will search again for any array (not just an array created for this
1261 host) and will assemble each assuming
1262 .BR \-\-update=homehost .
1263 This will change the host tag in the superblock so that on the next run,
1264 these arrays will be found without the second pass. The intention of
1265 this feature is to support transitioning a set of md arrays to using
1268 The reason for requiring arrays to be tagged with the homehost for
1269 auto assembly is to guard against problems that can arise when moving
1270 devices from one host to another.
1280 .BI \-\-raid\-devices= Z
1284 This usage is similar to
1286 The difference is that it creates an array without a superblock. With
1287 these arrays there is no difference between initially creating the array and
1288 subsequently assembling the array, except that hopefully there is useful
1289 data there in the second case.
1291 The level may raid0, linear, multipath, or faulty, or one of their
1292 synonyms. All devices must be listed and the array will be started
1304 .BI \-\-raid\-devices= Z
1308 This usage will initialise a new md array, associate some devices with
1309 it, and activate the array.
1313 option is given (as described in more detail in the section on
1314 Assemble mode), then the md device will be created with a suitable
1315 device number if necessary.
1317 As devices are added, they are checked to see if they contain raid
1318 superblocks or filesystems. They are also checked to see if the variance in
1319 device size exceeds 1%.
1321 If any discrepancy is found, the array will not automatically be run, though
1324 can override this caution.
1326 To create a "degraded" array in which some devices are missing, simply
1327 give the word "\fBmissing\fP"
1328 in place of a device name. This will cause
1330 to leave the corresponding slot in the array empty.
1331 For a RAID4 or RAID5 array at most one slot can be
1332 "\fBmissing\fP"; for a RAID6 array at most two slots.
1333 For a RAID1 array, only one real device needs to be given. All of the
1337 When creating a RAID5 array,
1339 will automatically create a degraded array with an extra spare drive.
1340 This is because building the spare into a degraded array is in general faster than resyncing
1341 the parity on a non-degraded, but not clean, array. This feature can
1342 be overridden with the
1346 When creating an array with version-1 metadata a name for the host is
1348 If this is not given with the
1352 will chose a name based on the last component of the name of the
1353 device being created. So if
1355 is being created, then the name
1360 is being created, then the name
1364 A new array will normally get a randomly assigned 128bit UUID which is
1365 very likely to be unique. If you have a specific need, you can choose
1366 a UUID for the array by giving the
1368 option. Be warned that creating two arrays with the same UUID is a
1369 recipe for disaster. Also, using
1371 when creating a v0.90 array will silently override any
1376 '''option is given, it is not necessary to list any component-devices in this command.
1377 '''They can be added later, before a
1381 '''is given, the apparent size of the smallest drive given is used.
1383 The General Management options that are valid with
1388 insist on running the array even if some devices look like they might
1393 start the array readonly \(em not supported yet.
1401 .I options... devices...
1404 This usage will allow individual devices in an array to be failed,
1405 removed or added. It is possible to perform multiple operations with
1406 on command. For example:
1408 .B " mdadm /dev/md0 \-f /dev/hda1 \-r /dev/hda1 \-a /dev/hda1"
1414 and will then remove it from the array and finally add it back
1415 in as a spare. However only one md array can be affected by a single
1426 MISC mode includes a number of distinct operations that
1427 operate on distinct devices. The operations are:
1430 The device is examined to see if it is
1431 (1) an active md array, or
1432 (2) a component of an md array.
1433 The information discovered is reported.
1437 The device should be an active md device.
1439 will display a detailed description of the array.
1443 will cause the output to be less detailed and the format to be
1444 suitable for inclusion in
1445 .BR /etc/mdadm.conf .
1448 will normally be 0 unless
1450 failed to get useful information about the device(s). However if the
1452 option is given, then the exit status will be:
1456 The array is functioning normally.
1459 The array has at least one failed device.
1462 The array has multiple failed devices such that it is unusable.
1465 There was an error while trying to get information about the device.
1470 The device should be a component of an md array.
1472 will read the md superblock of the device and display the contents.
1477 then multiple devices that are components of the one array
1478 are grouped together and reported in a single entry suitable
1480 .BR /etc/mdadm.conf .
1484 without listing any devices will cause all devices listed in the
1485 config file to be examined.
1489 The devices should be active md arrays which will be deactivated, as
1490 long as they are not currently in use.
1494 This will fully activate a partially assembled md array.
1498 This will mark an active array as read-only, providing that it is
1499 not currently being used.
1505 array back to being read/write.
1509 For all operations except
1512 will cause the operation to be applied to all arrays listed in
1517 causes all devices listed in the config file to be examined.
1524 .B mdadm \-\-monitor
1525 .I options... devices...
1530 to periodically poll a number of md arrays and to report on any events
1533 will never exit once it decides that there are arrays to be checked,
1534 so it should normally be run in the background.
1536 As well as reporting events,
1538 may move a spare drive from one array to another if they are in the
1541 and if the destination array has a failed drive but no spares.
1543 If any devices are listed on the command line,
1545 will only monitor those devices. Otherwise all arrays listed in the
1546 configuration file will be monitored. Further, if
1548 is given, then any other md devices that appear in
1550 will also be monitored.
1552 The result of monitoring the arrays is the generation of events.
1553 These events are passed to a separate program (if specified) and may
1554 be mailed to a given E-mail address.
1556 When passing event to program, the program is run once for each event
1557 and is given 2 or 3 command-line arguments. The first is the
1558 name of the event (see below). The second is the name of the
1559 md device which is affected, and the third is the name of a related
1560 device if relevant, such as a component device that has failed.
1564 is given, then a program or an E-mail address must be specified on the
1565 command line or in the config file. If neither are available, then
1567 will not monitor anything.
1571 will continue monitoring as long as something was found to monitor. If
1572 no program or email is given, then each event is reported to
1575 The different events are:
1579 .B DeviceDisappeared
1580 An md array which previously was configured appears to no longer be
1581 configured. (syslog priority: Critical)
1585 was told to monitor an array which is RAID0 or Linear, then it will
1587 .B DeviceDisappeared
1588 with the extra information
1590 This is because RAID0 and Linear do not support the device-failed,
1591 hot-spare and resync operations which are monitored.
1595 An md array started reconstruction. (syslog priority: Warning)
1601 is 20, 40, 60, or 80, this indicates that rebuild has passed that many
1602 percentage of the total. (syslog priority: Warning)
1606 An md array that was rebuilding, isn't any more, either because it
1607 finished normally or was aborted. (syslog priority: Warning)
1611 An active component device of an array has been marked as
1612 faulty. (syslog priority: Critical)
1616 A spare component device which was being rebuilt to replace a faulty
1617 device has failed. (syslog priority: Critial)
1621 A spare component device which was being rebuilt to replace a faulty
1622 device has been successfully rebuilt and has been made active.
1623 (syslog priority: Info)
1627 A new md array has been detected in the
1629 file. (syslog priority: Info)
1633 A newly noticed array appears to be degraded. This message is not
1636 notices a drive failure which causes degradation, but only when
1638 notices that an array is degraded when it first sees the array.
1639 (syslog priority: Critial)
1643 A spare drive has been moved from one array in a
1645 to another to allow a failed drive to be replaced.
1646 (syslog priority: Info)
1652 has been told, via the config file, that an array should have a certain
1653 number of spare devices, and
1655 detects that it has fewer that this number when it first sees the
1656 array, it will report a
1659 (syslog priority: Warning)
1663 An array was found at startup, and the
1666 (syslog priority: Info)
1676 cause Email to be sent. All events cause the program to be run.
1677 The program is run with two or three arguments, they being the event
1678 name, the array device and possibly a second device.
1680 Each event has an associated array device (e.g.
1682 and possibly a second device. For
1687 the second device is the relevant component device.
1690 the second device is the array that the spare was moved from.
1694 to move spares from one array to another, the different arrays need to
1695 be labelled with the same
1697 in the configuration file. The
1699 name can be any string. It is only necessary that different spare
1700 groups use different names.
1704 detects that an array which is in a spare group has fewer active
1705 devices than necessary for the complete array, and has no spare
1706 devices, it will look for another array in the same spare group that
1707 has a full complement of working drive and a spare. It will then
1708 attempt to remove the spare from the second drive and add it to the
1710 If the removal succeeds but the adding fails, then it is added back to
1714 The GROW mode is used for changing the size or shape of an active
1716 For this to work, the kernel must support the necessary change.
1717 Various types of growth are being added during 2.6 development,
1718 including restructuring a raid5 array to have more active devices.
1720 Currently the only support available is to
1722 change the "size" attribute
1723 for RAID1, RAID5 and RAID6.
1725 increase the "raid-disks" attribute of RAID1, RAID5, and RAID6.
1727 add a write-intent bitmap to any array which support these bitmaps, or
1728 remove a write-intent bitmap from such an array.
1732 Normally when an array is built the "size" it taken from the smallest
1733 of the drives. If all the small drives in an arrays are, one at a
1734 time, removed and replaced with larger drives, then you could have an
1735 array of large drives with only a small amount used. In this
1736 situation, changing the "size" with "GROW" mode will allow the extra
1737 space to start being used. If the size is increased in this way, a
1738 "resync" process will start to make sure the new parts of the array
1741 Note that when an array changes size, any filesystem that may be
1742 stored in the array will not automatically grow to use the space. The
1743 filesystem will need to be explicitly told to use the extra space.
1745 .SS RAID-DEVICES CHANGES
1747 A RAID1 array can work with any number of devices from 1 upwards
1748 (though 1 is not very useful). There may be times which you want to
1749 increase or decrease the number of active devices. Note that this is
1750 different to hot-add or hot-remove which changes the number of
1753 When reducing the number of devices in a RAID1 array, the slots which
1754 are to be removed from the array must already be vacant. That is, the
1755 devices that which were in those slots must be failed and removed.
1757 When the number of devices is increased, any hot spares that are
1758 present will be activated immediately.
1760 Increasing the number of active devices in a RAID5 is much more
1761 effort. Every block in the array will need to be read and written
1762 back to a new location. From 2.6.17, the Linux Kernel is able to do
1763 this safely, including restart and interrupted "reshape".
1765 When relocating the first few stripes on a raid5, it is not possible
1766 to keep the data on disk completely consistent and crash-proof. To
1767 provide the required safety, mdadm disables writes to the array while
1768 this "critical section" is reshaped, and takes a backup of the data
1769 that is in that section. This backup is normally stored in any spare
1770 devices that the array has, however it can also be stored in a
1771 separate file specified with the
1773 option. If this option is used, and the system does crash during the
1774 critical period, the same file must be passed to
1776 to restore the backup and reassemble the array.
1780 A write-intent bitmap can be added to, or removed from, an active
1781 array. Either internal bitmaps, or bitmaps stored in a separate file
1782 can be added. Note that if you add a bitmap stored in a file which is
1783 in a filesystem that is on the raid array being affected, the system
1784 will deadlock. The bitmap must be on a separate filesystem.
1786 .SH INCREMENTAL MODE
1790 .B mdadm \-\-incremental
1796 .B mdadm \-\-incremental \-\-rebuild
1799 .B mdadm \-\-incremental \-\-run \-\-scan
1803 This mode is designed to be used in conjunction with a device
1804 discovery system. As devices are found in a system, they can be
1806 .B "mdadm \-\-incremental"
1807 to be conditionally added to an appropriate array.
1810 performs a number of tests to determine if the device is part of an
1811 array, and which array is should be part of. If an appropriate array
1812 is found, or can be created,
1814 adds the device to the array and conditionally starts the array.
1818 will only add devices to an array which were previously working
1819 (active or spare) parts of that array. It does not currently support
1820 automatic inclusion of a new drive as a spare in some array.
1822 .B "mdadm \-\-incremental"
1823 requires a bug present in all kernels through 2.6.19, to be fixed.
1824 Hopefully this will be fixed in 2.6.20. Alternately apply the patch
1825 which is included with the mdadm source distribution. If
1827 detects that this bug is present, it will abort any attempt to use
1828 .BR \-\-incremental .
1832 makes are as follow:
1834 Is the device permitted by
1836 That is, is it listed in a
1838 line in that file. If
1840 is absent then the default it to allow any device. Similar if
1842 contains the special word
1844 then any device is allowed. Otherwise the device name given to
1846 must match one of the names or patterns in a
1851 Does the device have a valid md superblock. If a specific metadata
1852 version is request with
1856 then only that style of metadata is accepted, otherwise
1858 finds any known version of metadata. If no
1860 metadata is found, the device is rejected.
1863 Does the metadata match an expected array?
1864 The metadata can match in two ways. Either there is an array listed
1867 which identifies the array (either by UUID, by name, by device list,
1868 or by minor-number), the array was created with a
1872 matches that which is given in
1874 or on the command line.
1877 is not able to positively identify the array as belonging to the
1878 current host, the device will be rejected.
1882 keeps a list of arrays that is has partly assembled in
1883 .B /var/run/mdadm/map
1885 .B /var/run/mdadm.map
1886 if the directory doesn't exist). If no array exists which matches
1887 the metadata on the new device,
1889 must choose a device name and unit number. It does this based on any
1892 or any name information stored in the metadata. If this name
1893 suggests a unit number, that number will be used, otherwise a free
1894 unit number will be chosen. Normally
1896 will prefer to create a partitionable array, however if the
1900 suggests that a non-partitionable array is preferred, that will be
1904 Once an appropriate array is found or created and the device is added,
1906 must decide if the array is ready to be started. It will
1907 normally compare the number of available (non-spare) devices to the
1908 number of devices that the metadata suggests need to be active. If
1909 there are at least that many, the array will be started. This means
1910 that if any devices are missing the array will not be restarted.
1916 in which case the array will be run as soon as there are enough
1917 devices present for the data to be accessible. For a raid1, that
1918 means one device will start the array. For a clean raid5, the array
1919 will be started as soon as all but one drive is present.
1921 Note that neither of these approaches is really ideal. If it is can
1922 be known that all device discovery has completed, then
1926 can be run which will try to start all arrays that are being
1927 incrementally assembled. They are started in "read-auto" mode in
1928 which they are read-only until the first write request. This means
1929 that no metadata updates are made and no attempt at resync or recovery
1930 happens. Further devices that are found before the first write can
1931 still be added safely.
1935 .B " mdadm \-\-query /dev/name-of-device"
1937 This will find out if a given device is a raid array, or is part of
1938 one, and will provide brief information about the device.
1940 .B " mdadm \-\-assemble \-\-scan"
1942 This will assemble and start all arrays listed in the standard config file
1943 file. This command will typically go in a system startup file.
1945 .B " mdadm \-\-stop \-\-scan"
1947 This will shut down all array that can be shut down (i.e. are not
1948 currently in use). This will typically go in a system shutdown script.
1950 .B " mdadm \-\-follow \-\-scan \-\-delay=120"
1952 If (and only if) there is an Email address or program given in the
1953 standard config file, then
1954 monitor the status of all arrays listed in that file by
1955 polling them ever 2 minutes.
1957 .B " mdadm \-\-create /dev/md0 \-\-level=1 \-\-raid\-devices=2 /dev/hd[ac]1"
1959 Create /dev/md0 as a RAID1 array consisting of /dev/hda1 and /dev/hdc1.
1962 .B " echo 'DEVICE /dev/hd*[0\-9] /dev/sd*[0\-9]' > mdadm.conf"
1964 .B " mdadm \-\-detail \-\-scan >> mdadm.conf"
1966 This will create a prototype config file that describes currently
1967 active arrays that are known to be made from partitions of IDE or SCSI drives.
1968 This file should be reviewed before being used as it may
1969 contain unwanted detail.
1971 .B " echo 'DEVICE /dev/hd[a\-z] /dev/sd*[a\-z]' > mdadm.conf"
1973 .B " mdadm \-\-examine \-\-scan \-\-config=mdadm.conf >> mdadm.conf"
1975 This will find what arrays could be assembled from existing IDE and
1976 SCSI whole drives (not partitions) and store the information is the
1977 format of a config file.
1978 This file is very likely to contain unwanted detail, particularly
1981 entries. It should be reviewed and edited before being used as an
1984 .B " mdadm \-\-examine \-\-brief \-\-scan \-\-config=partitions"
1986 .B " mdadm \-Ebsc partitions"
1988 Create a list of devices by reading
1989 .BR /proc/partitions ,
1990 scan these for RAID superblocks, and printout a brief listing of all
1993 .B " mdadm \-Ac partitions \-m 0 /dev/md0"
1995 Scan all partitions and devices listed in
1996 .BR /proc/partitions
1999 out of all such devices with a RAID superblock with a minor number of 0.
2001 .B " mdadm \-\-monitor \-\-scan \-\-daemonise > /var/run/mdadm"
2003 If config file contains a mail address or alert program, run mdadm in
2004 the background in monitor mode monitoring all md devices. Also write
2005 pid of mdadm daemon to
2006 .BR /var/run/mdadm .
2008 .B " mdadm \-Iq /dev/somedevice"
2010 Try to incorporate newly discovered device into some array as
2013 .B " mdadm \-\-incremental \-\-rebuild \-\-run \-\-scan"
2015 Rebuild the array map from any current arrays, and then start any that
2018 .B " mdadm /dev/md4 --fail detached --remove detached"
2020 Any devices which are components of /dev/md4 will be marked as faulty
2021 and then remove from the array.
2023 .B " mdadm \-\-create \-\-help"
2025 Provide help about the Create mode.
2027 .B " mdadm \-\-config \-\-help"
2029 Provide help about the format of the config file.
2031 .B " mdadm \-\-help"
2033 Provide general help.
2044 lists all active md devices with information about them.
2046 uses this to find arrays when
2048 is given in Misc mode, and to monitor array reconstruction
2054 The config file lists which devices may be scanned to see if
2055 they contain MD super block, and gives identifying information
2056 (e.g. UUID) about known MD arrays. See
2060 .SS /var/run/mdadm/map
2063 mode is used. this file gets a list of arrays currently being created.
2066 does not exist as a directory, then
2067 .B /var/run/mdadm.map
2072 While entries in the /dev directory can have any format you like,
2074 has an understanding of 'standard' formats which it uses to guide its
2075 behaviour when creating device files via the
2079 The standard names for non-partitioned arrays (the only sort of md
2080 array available in 2.4 and earlier) either of
2086 where NN is a number.
2087 The standard names for partitionable arrays (as available from 2.6
2094 Partition numbers should be indicated by added "pMM" to these, thus "/dev/md/d1p2".
2098 was previously known as
2102 is completely separate from the
2104 package, and does not use the
2106 configuration file at all.
2109 For information on the various levels of
2113 .UR http://ostenfeld.dk/~jakob/Software\-RAID.HOWTO/
2114 http://ostenfeld.dk/~jakob/Software\-RAID.HOWTO/
2117 '''for new releases of the RAID driver check out:
2120 '''.UR ftp://ftp.kernel.org/pub/linux/kernel/people/mingo/raid-patches
2121 '''ftp://ftp.kernel.org/pub/linux/kernel/people/mingo/raid-patches
2126 '''.UR http://www.cse.unsw.edu.au/~neilb/patches/linux-stable/
2127 '''http://www.cse.unsw.edu.au/~neilb/patches/linux-stable/
2130 The latest version of
2132 should always be available from
2134 .UR http://www.kernel.org/pub/linux/utils/raid/mdadm/
2135 http://www.kernel.org/pub/linux/utils/raid/mdadm/